k8s-rdma-shared-dev-plugin
k8s-rdma-shared-dev-plugin copied to clipboard
"missing RDMA device spec for device 0000:e5:00.1, RDMA device \"issm\" not found"
Getting "missing RDMA device spec for device 0000:e5:00.1, RDMA device "issm" not found" error while creating a pod after installing k8s-rdma-shared-dev-plugin, what is the solution for this error.
You can run these command.
mst start
mst status -v
maybe the 0000:e5:00.1 id is ethernet interface card.
I also had the same problem. My logs are as follows:
error creating new device: "missing RDMA device spec for device 0000:e1:00.0, RDMA device \"issm\" not found"
I add the device's deviceID and vendors to RDMA Shared Device Plugin Configurations,then apply and restart the pods.The problem was solved.
I also encountered the same issue. My logs were as follows:
error creating new device: "missing RDMA device spec for device 0000:e1:00.0, RDMA device \"issm\" not found"
I added the device's deviceID and vendor information to the RDMA Shared Device Plugin configurations, then applied the changes and restarted the pods. The problem was resolved.
My RDMA Shared Device Plugin Configuration is as follows:
apiVersion: v1
kind: ConfigMap
metadata:
name: rdma-devices
namespace: kube-system
data:
config.json: |
{
"periodicUpdateInterval": 300,
"configList": [{
"resourceName": "cx5_bond_shared_devices_a",
"rdmaHcaMax": 1000,
"selectors": {
"vendors": ["15b3"],
"deviceIDs": ["1017","1019"]
}
},
{
"resourceName": "cx6dx_shared_devices_b",
"rdmaHcaMax": 500,
"selectors": {
"vendors": ["15b3"],
"deviceIDs": ["101d"]
}
}
]
}
assuming 0000:e5:00.1 belongs to an MLNX NIC, the error "issm" not found" means that some linux char device ( found under /dev/infiniband/issm<N>) is missing for the selected NIC. that means not all rdma modules were loaded. do you have rmda-core package installed ? it sets up udev rules to bind the needed drivers to MLNX NIC.
assuming 0000:e5:00.1 belongs to an MLNX NIC, the error "issm" not found" means that some linux char device ( found under /dev/infiniband/issm<N>) is missing for the selected NIC. that means not all rdma modules were loaded. do you have rmda-core package installed ? it sets up udev rules to bind the needed drivers to MLNX NIC.
That solved it for me! Thanks.